The PAX5 oncogene is expressed in N-type neuroblastoma cells and increases
tumorigenicity of a S-type cell line
Florian B.Baumann Kubetzko
1, Claudio di Paolo
1,
Charlotte Maag
1, Roland Meier
3, Beat W.Schafer
4,
David R.Betts
4, Rolf A.Stahel
1and Andreas
Himmelmann
1,2,51Forschungslabor Molekulare Onkologie, Klinik und Poliklinik fur Onkologie, Universitaetsspital Zurich, Haeldeliweg 4, CH-8044 Zurich, Switzerland,2Medizinische Klinik B, Universitaetsspital Zurich, Raemistrasse 100, CH-8091 Zurich, Switzerland,3Medizinische Klinik, Universitaets-Kinderklinik, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland and 4Abteilung fur Onkologie, Universitaets-Kinderklinik, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland
5To whom correspondence should be addressed Email: andreas.himmelmann@usz.ch
Neuroblastoma is a neural crest-derived neoplasm of
infancy with poor outcome in patients with advanced
disease. The oncogenic transcription factor PAX5 is an
important developmental regulator and is implicated in the
pathogenesis of several malignancies. Screening of
neuro-blastoma cell lines revealed PAX5 expression in a
malig-nant subset of neuroblastoma cells, so-called `N-type' cells,
but not in the more benign `S-type' neuroblastoma cells.
PAX5 expression was also detected in small cell lung
cancer, an aggressive tumor of neural crest origin. Based
on this observation we hypothesized that there could be a
relationship between PAX5 expression and the more
malig-nant phenotype of N-type cells. Stable PAX5 expression
was established in several clones of the S-type cell line
CA-2E. A noticeable difference in morphology of these
transfectants was observed and there was also a significant
increase in the proliferation rate. Moreover, PAX5
expressing clones gained the ability to form colonies in a soft
agar assay, a marker of tumorigenicity. Down-regulation
of PAX5 in several N-type cell lines and one small cell lung
cancer cell line utilizing small interfering RNA resulted in
a significant decrease in growth rate. Taken together we
propose PAX5 as an important factor for the maintenance
of the proliferative and tumorigenic phenotype of
neuro-blastoma. Our data, together with a recent study on the
role of PAX genes in cancer suggest that PAX5 and other
PAX transcription factors might be valuable targets for
cancer therapy.
Introduction
Neuroblastoma is the most common extracranial solid tumor in
children and is the leading cause of cancer death in children
between 1 and 4 years (1,2). It is a neoplasm of the neural crest,
a transient embryonic structure giving rise to a variety of
tissues (e.g. melanocytes, neurons, neuroendocrine tissue and
smooth muscle cells) (3). Histologically neuroblastoma tumors
are composed of distinct malignant cell types, including
neu-roblastic and stromal cells (1).
Cellular diversity is also characteristic of cell lines
estab-lished from explanted neuroblastoma tumors. A number of
sub-populations have been identified in these cultures, including
neuroblastic (N-type), substrate adherent (S-type) cells and
cells with an intermediate phenotype (I-type). These cell
types can be distinguished by a characteristic morphology
(4--6). Analysis of homogeneous cultures of these subtypes
has shown that they can also be distinguished by the expression
of a number of differentiation-associated markers. N-type cells
express neurotransmitter biosynthetic enzymes, chromogranin
A and neurofilament proteins while S-type cells synthesize
collagen, fibronectin and vimentin (4,7). In contrast to S-type
cells, N-type cells are tumorigenic in nude mice and form
colonies in soft agar (8). The difference in malignant
proper-ties between S- and N-type cell subtypes may have clinical
relevance (9). At the transcriptional level several differences in
gene expression have been detected. For example, BCL2 and
MYCN are usually expressed at a higher level in N-type cells
(10,11). The difference in MYCN expression was observed in
both MYCN amplified and non-amplified cell lines (11,12).
Recently it has been demonstrated that S-type in contrast to
N-type cells are more susceptible to TRAIL-mediated
apopto-sis due to their expression of caspase-8 (13,14).
PAX5 is a member of the PAX gene family of
developmen-tally regulated transcription factors that play a fundamental
role in organogenesis (15). All PAX transcription factors share
a DNA binding domainÐthe paired box. Their importance in
development has been underscored by several loss-of-function
mutations that cause lack of a specific structure or organ in
which the corresponding PAX gene is normally expressed.
Typically, PAX protein expression is down-regulated in the
adult organism. During mouse brain development Pax5 is
expressed predominantly at the midbrain--hindbrain boundary.
A weaker expression is found along the entire neural tube,
particularly in the ventricular zone of both alar and basal plates
(16). A similar expression pattern has been described recently
in human embryos (17). In addition, Pax5 is expressed during
all stages of B-cell development, except in plasma cells (18).
Pax5
ÿ=ÿmice lack mature B cells and show defects of the
inferior colliculus and anterior cerebellum (19).
In addition to the normal physiological function during
development, PAX transcription factors have been shown to
play an important role in tumorigenesis (reviewed in ref. 20).
First, a number of PAX genes are located at recurring,
tumor-specific chromosomal translocations, suggesting that PAX
genes have an oncogenic capacity when inappropriately
expressed. Secondly, several PAX genes are re-expressed in
malignant neoplasms, usually in tumors derived from tissue in
which the respective PAX gene is expressed during
develop-ment. For example, deregulated expression of PAX5 has been
described in medulloblastoma and lymphomas (21,22) and of
Abbreviations: MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer; scRNA, scrambled control; siRNA, small interfering RNA.
PAX3 in melanoma (23). It is speculated that the re-expressed
PAX genes promote tumor development and progression
by increasing proliferation and motility, while inhibiting
apoptosis (20).
Here we describe that PAX5 is expressed in aggressive
N-type neuroblastoma cell lines, while no expression was
detected in S-type cells. Over-expression of PAX5 in the
S-type cell line CA-2E restored several malignant properties,
in particular anchorage-independent growth. In addition
down-regulation of PAX5 in several N-type cell lines significantly
reduced their proliferation rate. These results provide
addi-tional evidence for an important role of the PAX gene family
in tumor development and suggest a role for PAX5 in
neuro-blastoma disease progression.
Materials and methods
Cell lines
ACN, CA-2E, BE(2)C, BE(2)M17, Imr32, Kelly, LAN-1, LAN-5, SH-310, SH-EP, SH-IN, SK-N-BE(2), SH-SY5Y and WSN were kindly provided by Nicole Gross (University Hospital, Lausanne, Switzerland). U1285 was a kind gift of Larisa Belyanskaya (Universitaetsspital, Zurich, Switzerland), MMLI157 was provided by Juerg Zarn (formerly Universitaetsspital Zurich, Switzerland) and Kis-1 was kindly provided by Hitoshi Ohno (Kyoto Univer-sity, Kyoto, Japan). SW2, OH1 and OH3 have been described previously (24). A549, Calu1, Calu3, Calu6, H460, NCI-H69 and NCI-H125 were obtained from ATCC (Rockville, MD). All cell lines were cultured in RPMI1640 supplemented with 10% FCS, 2 mML-glutamine, 50 IU/ml penicillin and 50 mg/ml streptomycin and maintained at 37C in a humified atmosphere with 5% CO2.
RT--PCR
RNA was isolated from cell lines using the RNeasy Mini Kit according to the manufacturer's instructions (Qiagen, Basel, Switzerland). Five micro-grams of total RNA were used to synthesize cDNA using the M-MuLV reverse transcriptase according to the manufacurer's instructions (MBI Fermentas, Vilnius, Lithuania). One-hundredth of the product was used for amplification using primers for PAX5 or GAPDH and PCR conditions as described elsewhere (25).
Plasmid construction
A mammalian PAX5 cDNA expression plasmid designated pX-13 was gener-ated by cloning a 3.3 kb XbaI fragment from phBSAP-1s (a kind gift of Meinrad Busslinger, Research Institute of Molecular Pathology, Vienna, Austria) into the mammalian expression vector pcDNA3.1/Hygro() (Invitrogen, Basel, Switzerland) using standard methods (26). This fragment contained the full-length PAX5 coding sequence (1176 bp) together with 77 bp of the 50-UTR and 2026 bp of the 30-UTR sequence. The correct orientation of the PAX5-cDNA was confirmed by sequencing (Microsynth, Balgach, Switzerland).
Cell transfection and generation of stable cell lines
Transfection of the plasmids was performed with Superfect following the manufacturer's instructions (Qiagen, Basel, Switzerland). Stable clones were generated by transfecting CA-2E cells with plasmids pX-13 or pcDNA3.1/ Hygro() for 20 h. The transfection medium was aspirated and cells were cul-tured in fresh medium for another day before adding 400 mg/ml hygromycin B for selection. Single clones were achieved by plating the cells in 96-well plates after 7--10 days of selection at a density of 0.5 cells/well in 100 ml selective medium. Single clones were expanded and screened for PAX5 expression by immunoblotting.
Cell lysis and immunoblotting
Cells were harvested and pellets were frozen at ÿ80C for at least 1 h. RIPA buffer was used for cell lysis as described (22). Between 80 and 120 mg of total lysate was separated by SDS--PAGE on a 10% gel. Transfer and immunoblotting conditions have been described previously (24). For PAX5 and actin detection the Pax5-C14 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) or an anti-Actin antibody (ICN Biomedicals GmbH, Eschwege, Germany) have been used respectively and were combined with the corre-sponding secondary antibodies. Chemoluminescent signals were detected using the ECL Kit (Amersham Biosciences, Buckinghamshire, UK). Band intensities were measured using standard methods.
Growth experiments
5000 cells were plated in 6-well plates containing 3 ml RPMI 1640 supple-mented with 5% FCS, 2 mML-glutamine, 50 IU/ml penicillin and 50 mg/ml streptomycin. Cells were grown at standard conditions and harvested after 4 days. After centrifugation the cells were re-suspended in PBS and the number of cells was determined using a Coulter counter (Beckman Coulter, Fullerton, CA). The harvesting and counting procedure was repeated every 24 h for an additional 5 days. Experiments were performed in triplicate and repeated four times with similar results.
Soft agar assay
To determine the ability of anchorage-independent growth a soft agar clono-genic assay was performed with PAX5 positive and negative clones. A base solution consisting of 1.5 ml 0.5% agar in culture medium was poured in each well of a 6-well plate. 3000 cells were re-suspended in 1.5 ml top solution consisting of 0.35% agar in culture medium and subsequently seeded on top of the polymerized base solution. All solutions were kept at 40C before pouring to prevent early agar polymerization and to ensure survival of the cells. Plates were incubated at 37C under standard culture conditions for 28--32 days and colony formation was visualized by staining with 0.01% crystal violet and quantified by counting. All experiments were performed in triplicate and experiments were repeated five times with similar results.
Xenografts
Six to eight week old female CD-1 (ICR nu/nu) mice (Charles River Lab-oratories, Sulzberg, Germany) were used and kept under specific pathogen-free conditions according to the guidelines of the Veterinary Office of the Kanton Zurich, Switzerland. Ten million cells re-suspended in sterile PBS were injected subcutaneously on the flanks. Tumor growth was assessed by measuring tumor size once per week.
Immunohistochemistry
Formalin-fixed and paraffin-embedded tumor samples were subjected to immunohistochemical staining for PAX5 and Vimentin as described elsewhere (27). The primary anti-Pax-5 (Transduction Laboratories, Lexington, KY) and anti-Vimentin (DAKO, Glostrup, Denmark) were used.
Transfection with small interfering RNA (siRNA) and measurement of cell proliferation
Briefly, for each experiment 5 105cells/well were plated in 6-well plates and transfected with 80 nM siRNA or scrambled control RNA (scRNA) for 24 h using Lipofectamine 2000 as a transfection mediator according to the manu-facturer's instructions (Invitrogen, Basel, Switzerland). The sequences of the sense strand and of the complementary strand of the siRNA and scRNA were as follows: siRNA sense strand 50-CGGCCACUCGCUUCCGGGCTT-30, complementary strand 50-GCCCGGAAGCGAGUGGCCGTT-30 and scRNA sense strand 50-GCUCCGUCGACUGCGCGCC-30, complementary strand 50-GGCGCGCAGUCGACGGAGCTT-30. These compounds were ordered as double-stranded RNA with a 2-nt DNA overhang (i.e. TT) at both 30-ends (Proligo, Paris, France).
Cell proliferation was measured by use of a colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay as described previously (28). Twenty-four hours after treatment with siRNA or scRNA, cells were plated in 96-well plates and incubated in standard culture medium containing only 5% FCS under standard conditions. The amount of plated cells per well was as follows: LAN-1 and OH1 (1.5 104), Imr32 and MMLI157 (2 104). The first MTT assay was performed immediately after distributing the cells in order to confirm the accuracy of the plating. Cell proliferation was assessed 24, 48, 72 and 96 h after plating. Data recording and analysis was performed as described previously (28). An individual standard calibration curve was constructed for each cell line and absorption values were equalized with the cell number by use of these standard curves.
Results
Expression of PAX5 in N-type neuroblastoma and small cell
lung cancer cell lines
Based on the observation that deregulated expression of
PAX genes is often observed in tumors derived from tissues
in which the respective PAX gene is expressed during normal
organogenesis, we screened tumors of the nervous system for
PAX5 expression by RT--PCR. Expression of PAX5 mRNA
was detected in seven N-type and parental neuroblastoma
cell lines while no transcripts were present in all four S-type
neuroblastoma cell lines and in two out of three I-type cell
lines (Figure 1A). In order to compare the relative abundance
of the two PAX5 isoforms, which differ only in the first exon
(29), we performed a quantitative Real-Time PCR with primers
specific for isoforms 1A and 1B. All neuroblastoma cell lines
that were positive for PAX5 in the conventional RT--PCR
expressed isoform 1B. It was only in a few cases that a very
low expression of the 1A isoform could be detected as well
(unpublished data). These findings are consistent with data
showing that the 1B promoter is predominantly used for PAX5
expression in non-B lymphocytes (30). PAX5 expression in
N-type and absence in S-type cells could be confirmed at the
protein level by immunoblotting. A major band of the expected
size of ~50 kDa was detected in lysates prepared from N-type
cell lines and the B-cell line Kis-1, but not in S-type cells. The
minor bands above the band representing PAX5 are non-specific
and also visible in S-type cells. An actin control is included
to ensure equal loading (Figure 1B).
As neuroblastoma is a neural crest-derived tumor, we also
screened additional tumors of the neural crest for PAX5 mRNA
expression. While no expression could be detected in five
melanoma cell lines (unpublished data), expression of PAX5
was found in five small cell lung cancer (SCLC) cell lines but
not in five non-small cell lung cancer (NSCLC) cell lines,
which are of endodermal origin (Figure 1C).
In summary, PAX5 expression was detected at the mRNA
and protein level in N-type neuroblastoma and SCLC cell lines
but not in any of the S-type neuroblastoma, melanoma and
non-neural crest-derived NSCLC cell lines, suggesting a role
for PAX5 in the more aggressive phenotype of these cell lines.
Stable expression of PAX5 in S-type neuroblastoma cell line
CA-2E
To investigate the possibility that the differential expression of
PAX5 in N- but not S-type neuroblastoma cell lines could in part
account for their distinctive phenotypes, we generated several
independent clones of the S-type neuroblastoma cell line CA-2E
with stable expression of PAX5. To this end we used the
PAX5-cDNA expression vector pX-13 and the empty cloning vector
pcDNA3.1/Hygro() as a control. A number of clones
trans-fected with either vector pX-13 or the control vector were
isolated after transfection and limiting dilution. Chromosomal
integration of the PAX5 expression vector or the empty cloning
vector was confirmed by vector-specific primers. PAX5
expres-sion in these clones was determined by immunoblotting
(Figure 2A). Three PAX5 expressing clones (P-16, P-25 and
P-28) and three clones transfected with the empty vector (C-43,
C-45 and C-50) were randomly selected for further phenotypic
analysis. The use of a control vector transfected in parallel and
the analysis of three independent clones of each transfected
vector should eliminate the possibility of selecting pre-existing
clones. To exclude this possibility using an additional method,
we performed fluorescence in situ hybridization (FISH)
experi-ments with a MYCN probe (Vysis, Downers Grove, IL). This
experiment showed the presence of two major populations with
either three or four MYCN copies per cell in the untransfected
CA-2E cell line. There was no significant difference in the
number of cells with three or four copies between the P and
C clones, thereby excluding selection of a subclone with a
particular MYCN copy number (unpublished data).
PAX5 expression in CA-2E neuroblastoma cells changes cell
morphology and increases growth rate
Direct microscopy of cell cultures of PAX5 expressing CA-2E
cells revealed distinct morphologic differences compared with
control cells. The latter consisted of flat, epithelial-like cells
with a large cytoplasm with no significant difference to
untransfected CA-2E cells (Figure 2B). In contrast, PAX5
expressing cells became smaller, rounder and were less
attached to plastic which was closer in morphology to N-type
cells (Figure 2C). The observed cell rounding was not due
to apoptosis as shown by propidium iodine staining followed
by FACS analysis (unpublished data).
Growth kinetics were analyzed by seeding 5000 cells of each
of two positive and negative PAX5 clones and counting cell
numbers after different time points. Compared with PAX5
negative clones, the positive clones grew significantly faster
Fig. 1. Expression of PAX5 in neuroblastoma and lung cancer cell lines. Qualitative analysis of PAX5 (upper panels) and GAPDH (lower panels) expression using RT--PCR with cDNA produced as described. (A) S-type neuroblastoma cell ines SH-EP, CA-2E, SH-310 and WSN (lanes 1--4), I-type cell lines SH-IN, ACN and BE(2)C (lanes 5--7), N-type and parental cell lines LAN-1, SH-SY5Y, Imr32, Kelly, BE(2)M17, LAN-5, SK-N-BE(2) (lanes 8--14) and the B cell lymphoma cell line Kis-1 without (lane 15) and with reverse transcriptase (lane 16) as a control. (B) Immunoblot analysis of PAX5 expression using 100 mg of total lysate per lane, for the Kis-1 cell line only 60 mg were loaded. The major band of ~50 kDa represents PAX5 (4), minor bands are non-specific. The actin control is also shown. N-type neuroblastoma cell lines LAN-1, Imr32, LAN-5, SH-SY5Y and SK-N-BE(2) (lanes 1--5) and S-type cell lines SH-EP and CA-2E (lanes 6--7). (C) RT--PCR with cDNA from the SCLC cell lines U1285, NCI-H69, SW2, OH1 and OH3 (lanes 1--5) and the NSCLC cell lines A549, NCI-H125, Calu1, Calu3 and Calu6 (lanes 6--10).
and to a higher cell number before the growth curves reached
their plateau (Figure 3). Moreover, there was no difference in
growth kinetics between the CA-2E cell line and the control
vector transfected cell line (unpublished data).
PAX5 expression greatly increases anchorage-independent
growth of CA-2E cells
Previous studies have shown that S-type cells isolated from
mixed neuroblastoma cell populations are less tumorigenic
than N-type cells, based on their differential ability for colony
formation in soft agar assays or tumor formation in nude mice
(8). Consistent with these observations CA-2E cells formed
nearly no colonies in our soft agar system, in contrast to the
N-type cell lines SH-SY5Y and LAN-1 (unpublished data).
PAX5 expressing CA-2E clones formed a large number of
colonies that were nevertheless smaller in size compared
with colonies formed by the N-type cell lines SH-SY5Y and
LAN-1. The control clones were comparable with the
wild-type CA-2E cells with almost no colony formation (Figure 4).
These observations could be reproduced for all clones in
several different experiments. Taken together these results
indicate that PAX5 expression can increase
anchorage-independent colony formation when over-expressed in the
appropriate cellular background.
PAX5 expression promotes tumor formation in nude mice
In in vivo experiments the PAX5 expressing P-25 and the
non-expressing C-43 cells were injected into four CD-1 nude mice
each on the lateral flank. Two mice injected with P-25
devel-oped small tumors with a size of 130 and 230 mm
3after
4 months whereas mice injected with C-43 did not form any
tumors. To confirm this result in a larger population we
injected P-25 cells into seven mice. Although we injected the
cells in both flanks only two mice developed tumors with a size
of 85 mm
3. Immunohistochemical analysis of the tumors
from both experiments confirmed that they were derived
from P-25 as they expressed PAX5 as well as the CA-2E
marker Vimentin (Figure 5A and B). Immunoreactivity for
PAX5 was present in the nucleus and co-localized with
Vimentin staining in the same cells.
PAX5 down-regulation with siRNA decreases cell
proliferation of N-type neuroblastoma cells
As shown in our growth experiments, stable PAX5 expression
significantly increased the growth rate of CA-2E cells. As
N-type cells already express endogenous PAX5 we exploited
the RNA interference (RNAi) technique (31) in order to
spe-cifically down-regulate PAX5 mRNA. We designed an siRNA
against PAX5 mRNA and an scRNA consisting of the same
Fig. 2. PAX5 expression induces morphological changes in CA-2E cells. The S-type CA-2E neuroblastoma cell line was transfected with the human PAX5 encoding expression vector pX-13 or the empty vector pcDNA3.1/ Hygro() as a control. Independent clones were isolated by limiting dilution and analyzed for PAX5 expression and morphology. (A) Immunoblot of 80 mg total cell lysates obtained from the three PAX5 expressing clones P-16, P-25 and P-28 (lanes 2, 4 and 6) and from three control clones C-43, C-45 and C-50 (lanes 1, 3 and 5) (upper panel). The equal amount of lysate was verified by a specific staining for actin (lower panel). Photomicrographs of cell cultures from the representative PAX5 expressing clone P-28 (C) and from the control clone C-45 (B) are also shown and a 50 mm scale is indicated. PAX5 expressing cells are smaller, rounder and less attached to the culture flask.
Fig. 3. Increased proliferation of PAX5 expressing CA-2E cells. 5000 cells of two PAX5 expressing (P-16 and P-25) and two non-expressing clones (C-43 and C-45) were plated and growth was recorded. Cell numbers were assessed with a Coulter counter at indicated time points. Cells were first counted on day 5, which is set as zero in the figure. The experiment was performed in triplicate for each clone and time point. Means and standard deviations are shown.
bases organized in a randomized combination without any
homology to human genomic sequences. The N-type
neuro-blastoma cell lines LAN-1 and Imr32 as well as the SCLC cell
line OH1 were either transfected with siRNA or the same
amount of scRNA. The melanoma cell line MMLI157, which
does not express endogenous PAX5 was used as a control.
Immunoblot analysis of cells transfected with siRNA and
har-vested 72 h after transfection showed a down-regulation of
the PAX5 protein level by 32% in LAN-1 cells and by 12% in
Imr32 and OH1 cells when compared with scRNA treatment
(Figure 6A--C). Staining for actin was used as a loading control
and is also shown.
Proliferation of transfected cells was measured by a MTT
assay. A modest but significant decrease in growth of the
siRNA transfected cells compared with scRNA transfected
cells was first observed 48 h after plating (Figure 6A--C).
The effect was greatest in LAN-1 where the PAX5
down-regulation worked most effective and small in OH1 where
PAC5 was weakly down-regulated. In addition the control
cell line MMLI157 was unaffected by either treatment (Figure
6D).
Discussion
In this study we report the expression of the oncogenic
transcription factor PAX5 in a subset of neuroblastoma cell
lines characterized by an aggressive phenotype, the so-called
`N-type' cells. In contrast, PAX5 was not expressed in any
of the tested S-type neuroblastoma cell lines. Further screening
of neural crest-derived tumors demonstrated expression in
SCLC but not in melanoma cell lines. The data show a distinct
expression pattern of PAX5 in neural crest-derived tumors with
a neuronal or neuroendocrine differentiation, indicating that
PAX5 might be a valuable marker for these derivatives. The
lack of expression in both melanoma cell lines and S-type cell
lines is consistent with the model of S-type cells being
differ-entiated along the melanocyte pathway (4).
More significantly, the expression of PAX5 in N-type
neuro-blastoma cells might be an important determinant of the more
malignant phenotype of these cells. To test this hypothesis two
approaches were taken. First, PAX5 was over-expressed in the
S-type neuroblastoma cell line CA-2E, resulting in a
remark-able phenotypic change. Secondly, PAX5 was down-regulated
in N-type cell lines and SCLC cell line by means of the siRNA
technique. A recent report has shown the coexistence of
subpopulations with different MYCN copy numbers in the
neuroblastoma cell line SK-N-SH and the derived cell lines
SH-SY5Y (N-type) and SH-EP (S-type). It was suggested that
the phenomenon of phenotypic interconversion is likely to
result from selection of pre-existing clones that can outgrow
Fig. 4. PAX5 expressing CA-2E cells are capable of growing anchorage-independently. The soft agar assay was performed with three PAX5 expressing (P-16, P-25 and P-28) and three PAX5 non-expressing clones (C-43, C-45 and C-50) derived from the CA-2E neuroblastoma cell line. (A) The data summarized in the table represent the means and standard deviations of a representative experiment carried out in triplicate. The lower panel shows pictures of the PAX5 expressing clone P-16 (B) and of the
non-expressing clone C-43 (C) after staining with 0.01% crystal violet. Fig. 5. Tumors established from the CA-2E transfected clone P-25 express markers PAX5 and Vimentin. Paraffin sections of a xenograft of the PAX5 expressing CA-2E clone P-25 established in a CD-1 nude mouse are shown. The tumor was isolated 4 months after injection into the lateral flank with a size of 230 mm3. Immunohistochemical stainings for PAX5 (A) and Vimentin (B) were performed as described. Magnification is 280-fold.
under certain conditions (32). The selection of a subclone with
a higher MYCN copy number could therefore account for the
increase in aggressiveness seen in our clones transfected with
PAX5. To exclude this important possibility we performed
FISH analysis with the MYCN probe in the untransfected
CA-2E cell line and the transfected clones. We could indeed
detect the coexistence of two populations with respect to
MYCN copy number, but no evidence of selection after
trans-fection. Therefore, these results provide evidence that the
phenotypic changes observed by us, including the increase of
tumorigenicity, are a result of the expression of PAX5 in
CA-2E cells.
The cellular background of the CA-2E cell line proved to be
particularly permissive to mediate the effects of PAX5
expres-sion, most likely because interacting co-factors required for
PAX5 function are present. CA-2E was established from the
bone marrow aspirate of a 16-month-old patient with
pro-gressive disease (33). The cells have the typical morphology
of S-type cells and do not form colonies in a soft agar assay.
PAX5 expression resulted in a striking increase in colony
numbers. The size of the individual colonies, however, was
smaller than those formed by the N-type cell line SH-SY5Y.
Consistent with this observation, PAX5 expressing cells
formed tumors only with low efficiency in animals. These
results are reminiscent of observations made when analysing
the transforming capacity of the PAX3--FKHR fusion protein
(34). Expression of this fusion protein in chicken embryo
fibroblasts led to cellular transformation in vitro, as shown
in focus and soft agar colony assays. In animal experiments,
however, no in vivo transforming capacity could be
demon-strated. It is possible that tumor induction in vivo requires
additional genetic changes or induction of additional genes
(e.g. responsible for neo-angiogenesis) that are not required
for colony formation in soft agar. In addition, several other
attempts to induce tumors in transgenic models by
over-expressing PAX5 or PAX3-FKHR have failed (35--37). These
Fig. 6. Down-regulation of PAX5 expression in N-type neuroblastoma and SCLC cell lines leads to a reduced growth rate. Endogenous PAX5 expression in the N-type neuroblastoma cell lines LAN-1 (A) and Imr32 (B) as well as in the SCLC cell line OH1 (C) was targeted by transfection of 80 nM of a sequence specific siRNA or 80 nM of a scrambled control scRNA. PAX5 protein level was determined 72 h after transfection by an immunoblot using 100 mg from total lysates (A--C). Loading of equal amounts was verified by immunoblotting for actin. The proliferation of transfected cells was assessed using the MTT colorimetric assay at indicated time points (A--C). MMLI157 was thereby used as a negative control (D). The number of cells was calculated according to a calibration curve to correlate optical density and cell number. An individual standard curve was constructed for each cell line. The data represent mean values and standard deviations of an experiment carried out in quintuplicate.
results could indicate that deregulated expression of PAX
genes is insufficient to cause tumor formation in vivo.
However, these results do not exclude an important role for
PAX gene expression in tumor progression or maintenance.
Our observation that targeting of PAX5 expression in N-type
neuroblastoma cell lines decreases proliferation supports this
hypothesis.
Our results confirm other observations implicating mouse
Pax5 as a positive regulator of cell proliferation (38). In
addi-tion down-regulaaddi-tion of PAX5 in N-type cells affected
growth, and we are currently trying to identify the target genes
mediating this effect. A novel finding is our observation that
PAX5 expression had a strong effect on S-type cell morphology
and adhesion. This suggests that PAX5 might also regulate
expression of cytoskeletal components and/or proteins of
extracellular matrix.
Our study is limited by the fact that the effect of PAX5
over-expression was studied in only one S-type cell line.
Never-theless, the expression data and the effects of PAX5 expression
in the S-type neuroblastoma cell line CA-2E suggest a role for
PAX5 in the molecular pathogenesis of neuroblastoma tumors.
Whether PAX5 might serve as a molecular target for therapy
is not known. To investigate this possibility we have started
an analysis of PAX5 expression in tumor samples. Recent
findings of widespread expression of PAX genes in numerous
common tumors and dependence of tumor viability on that
expression suggest that PAX5 and other PAX transcription
factors might be valuable targets for cancer therapy (39). It is
clear however, that the PAX5 transfected CA-2E cell line is
a very useful model system to study the oncogenic function of
PAX5. For example, the target genes mediating the oncogenic
effect of PAX5 can be identified using microarray technology.
Using this technique we have already identified several target
genes that are regulated by PAX5 also in cell lines other than
neuroblastoma cells (manuscript in preparation). The results
of these studies will provide answers on how PAX5 performs
its oncogenic function.
Acknowledgements
The authors are grateful to Thomas Stallmach who performed numerous immunohistochemical stainings of our mice tumor sections. We thank Sally Donaldson for her valuable scientific input and the critical reading of the manuscript and we thank Dario Neri for supporting this research project. This study was funded by a grant of the Swiss Cancer League to A.H.
References
1.Brodeur,G.M. (2003) Neuroblastoma: biological insights into a clinical enigma. Nature Rev. Cancer, 3, 203--216.
2.Spix,C., Aareleid,T., Stiller,C., Magnani,C., Kaatsch,P. and Michaelis,J. (2001) Survival of children with neuroblastoma. Time trends and regional differences in Europe, 1978--1992. Eur. J. Cancer, 37, 722--729. 3.Le Douarin,N. and Kalcheim,C. (1999) The Neural Crest. Cambridge
University Press, Cambridge.
4.Ciccarone,V., Spengler,B.A., Meyers,M.B., Biedler,J.L. and Ross,R.A. (1989) Phenotypic diversification in human neuroblastoma cells: expres-sion of distinct neural crest lineages. Cancer Res., 49, 219--225. 5.Thiele,C.J. (1999) Neuroblastoma. In Masters,J.R.W. and Palsson,B. (eds),
Human Cell Culture. Kluwer Academic Publishers, Dordrecht, The Netherlands, Vol. 1, pp. 21--53.
6.Ross,R.A., Spengler,B.A., Domenech,C., Porubcin,M., Rettig,W.J. and Biedler,J.L. (1995) Human neuroblastoma I-type cells are malignant neural crest stem cells. Cell Growth Differ., 6, 449--456.
7.Ross,R.A., Hein,A.M., Braca,J.A.,3rd, Spengler,B.A., Biedler,J.L. and Scammell,J.G. (2002) Glucocorticoids induce neuroendocrine cell
differentiation and increase expression of N-myc in N-type human neuroblastoma cells. Oncol. Res., 13, 87--94.
8.Biedler,J.L., Spengler,B.A.,T.,C. and Ross,R.A. (1988) Transdifferentiation of human neuroblastoma cells results in coordinate loss of neuronal and malignant properties. In Evans,A., D'Angio,G., Knudson,A. and Seeger,R. (eds), Advances in Neuroblastoma Research 2. Alan R. Liss, Inc., New York, pp. 265--276.
9.La Quaglia,M.P. and Manchester,K.M. (1996) A comparative analysis of neuroblastic and substrate-adherent human neuroblastoma cell lines. J. Pediatr. Surg., 31, 315--318.
10.Isaacs,J.S., Hardman,R., Carman,T.A., Barrett,J.C. and Weissman,B.E. (1998) Differential subcellular p53 localization and function in N- and S-type neuroblastoma cell lines. Cell Growth Differ., 9, 545--555. 11.Sadee,W.,Yu,V.C., Richards,M.L., Preis,P.N., Schwab,M.R., Brodsky,F.M.
and Biedler,J.L. (1987) Expression of neurotransmitter receptors and myc protooncogenes in subclones of a human neuroblastoma cell line. Cancer Res., 47, 5207--5212.
12.Foley,J., Cohn,S.L., Salwen,H.R., Chagnovich,D., Cowan,J., Mason,K.L. and Parysek,L.M. (1991) Differential expression of N-myc in pheno-typically distinct subclones of a human neuroblastoma cell line. Cancer Res., 51, 6338--6345.
13.Hopkins-Donaldson,S., Bodmer,J.L., Bourloud,K.B., Brognara,C.B., Tschopp,J. and Gross,N. (2000) Loss of caspase-8 expression in highly malignant human neuroblastoma cells correlates with resistance to tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis. Cancer Res., 60, 4315--4319.
14.Eggert,A., Grotzer,M.A., Zuzak,T.J., Wiewrodt,B.R., Ikegaki,N. and Brodeur,G.M. (2000) Resistance to TRAIL-induced apoptosis in neuro-blastoma cells correlates with a loss of caspase-8 expression. Med. Pediatr. Oncol., 35, 603--607.
15.Chi,N. and Epstein,J.A. (2002) Getting your Pax straight: Pax proteins in development and disease. Trends Genet., 18, 41--47.
16.Asano,M. and Gruss,P. (1992) Pax-5 is expressed at the midbrain-hindbrain boundary during mouse development. Mech. Dev., 39, 29--39. 17.Gerard,M., Abitbol,M., Delezoide,A.L., Dufier,J.L., Mallet,J. and
Vekemans,M. (1995) PAX-genes expression during human embryonic development, a preliminary report. C. R. Acad. Sci. III, 318, 57--66. 18.Adams,B., Dorfler,P., Aguzzi,A., Kozmik,Z., Urbanek,P., Maurer-Fogy,I.
and Busslinger,M. (1992) Pax-5 encodes the transcription factor BSAP and is expressed in B lymphocytes, the developing CNS and adult testis. Genes Dev., 6, 1589--1607.
19.Urbanek,P., Fetka,I., Meisler,M.H. and Busslinger,M. (1997) Cooperation of Pax2 and Pax5 in midbrain and cerebellum development. Proc. Natl Acad. Sci USA, 94, 5703--5708.
20.Schafer,B.W. (1998) Emerging roles for PAX transcription factors in cancer biology. Gen. Physiol. Biophys., 17, 211--224.
21.Kozmik,Z., Sure,U., Ruedi,D., Busslinger,M. and Aguzzi,A. (1995) Deregulated expression of PAX5 in medulloblastoma. Proc. Natl Acad. Sci. USA, 92, 5709--5713.
22.Krenacs,L., Himmelmann,A.W., Quintanilla-Martinez,L., Fest,T., Riva,A., Wellmann,A., Bagdi,E., Kehrl,J.H., Jaffe,E.S. and Raffeld,M. (1998) Transcription factor B-cell-specific activator protein (BSAP) is differen-tially expressed in B cells and in subsets of B-cell lymphomas. Blood, 92, 1308--1316.
23.Scholl,F.A., Kamarashev,J., Murmann,O.V., Geertsen,R., Dummer,R. and Schafer,B.W. (2001) PAX3 is expressed in human melanomas and contributes to tumor cell survival. Cancer Res., 61, 823--826.
24.Hopkins-Donaldson,S., Ziegler,A., Kurtz,S., Bigosch,C., Kandioler,D., Ludwig,C., Zangemeister-Wittke,U. and Stahel,R. (2003) Silencing of death receptor and caspase-8 expression in small cell lung carcinoma cell lines and tumors by DNA methylation. Cell Death Differ., 10, 356--364. 25.Adshead,J.M., Ogden,C.W., Penny,M.A., Stuart,E.T. and Kessling,A.M.
(1999) The expression of PAX5 in human transitional cell carcinoma of the bladder: relationship with de-differentiation. Br. J. Urol. Int., 83, 1039--1044.
26.Sambrook,J. and Russell,D.W. (2001) Molecular Cloning A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
27.Koesters,R., Niggli,F., von Knebel Doeberitz,M. and Stallmach,T. (2003) Nuclear accumulation of beta-catenin protein in Wilms' tumours. J. Pathol., 199, 68--76.
28.Olie,R.A., Hall,J., Natt,F., Stahel,R.A. and Zangemeister-Wittke,U. (2002) Analysis of ribosyl-modified, mixed backbone analogs of a bcl-2/bcl-xL antisense oligonucleotide. Biochim. Biophys. Acta, 1576, 101--109. 29.Busslinger,M., Klix,N., Pfeffer,P., Graninger,P.G. and Kozmik,Z. (1996)
locus adjacent to two alternative PAX-5 promoters in a diffuse large-cell lymphoma. Proc. Natl Acad. Sci. USA, 93, 6129--6134.
30.Morrison,A.M., Jager,U., Chott,A., Schebesta,M., Haas,O.A. and Busslinger,M. (1998) Deregulated PAX-5 transcription from a translo-cated IgH promoter in marginal zone lymphoma. Blood, 92, 3865--3878. 31.Fire,A., Xu,S., Montgomery,M.K., Kostas,S.A., Driver,S.E. and Mello,C.C.
(1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 391, 806--811.
32.Cohen,N., Betts,D.R., Rechavi,G., Amariglio,N. and Trakhtenbrot,L. (2003) Clonal expansion and not cell interconversion is the basis for the neuroblast and nonneuronal types of the SK-N-SH neuroblastoma cell line. Cancer Genet. Cytogenet., 143, 80--84.
33.Suardet,L., Gross,N., Gaide,A.C., Beck,D. and Eliason,J.F. (1989) Epidermal growth factor responsiveness of a new human neuroblastoma cell line. Int. J. Cancer, 44, 661--668.
34.Scheidler,S., Fredericks,W.J., Rauscher,F.J.,3rd, Barr,F.G. and Vogt,P.K. (1996) The hybrid PAX3-FKHR fusion protein of alveolar rhabdomyo-sarcoma transforms fibroblasts in culture. Proc. Natl Acad. Sci. USA, 93, 9805--9809.
35.Anderson,M.J., Shelton,G.D., Cavenee,W.K. and Arden,K.C. (2001) Embryonic expression of the tumor-associated PAX3-FKHR fusion pro-tein interferes with the developmental functions of Pax3. Proc. Natl Acad. Sci. USA, 98, 1589--1594.
36.Lagutina,I., Conway,S.J., Sublett,J. and Grosveld,G.C. (2002) Pax3-FKHR knock-in mice show developmental aberrations but do not develop tumors. Mol. Cell. Biol., 22, 7204--7216.
37.Steinbach,J.P., Kozmik,Z., Pfeffer,P. and Aguzzi,A. (2001) Overexpression of Pax5 is not sufficient for neoplastic transformation of mouse neuroectoderm. Int. J. Cancer, 93, 459--467.
38.Wakatsuki,Y., Neurath,M.F., Max,E.E. and Strober,W. (1994) The B cell-specific transcription factor BSAP regulates B cell proliferation. J. Exp. Med., 179, 1099--1108.
39.Muratovska,A., Zhou,C., He,S., Goodyer,P. and Eccles,M.R. (2003) Paired-box genes are frequently expressed in cancer and often required for cancer cell survival. Oncogene, 22, 6045--6053.